Organizations generally like to make the headlines with their success stories and not about the failure web sites due to unexpected user load. Industry study reveals that poor performing IT applications cost industrialized nations almost $45 billion annually. Nowadays most of the organizations understand the importance of performance testing and the impact of bypassing it. But most of the time, the assumptions made during the performance engineering phases become trumpery, leading to system failures because of the unexpected load. The root cause analysis of various applications in our organization revealed that defect leakage has occurred during the workload modeling phase of the performance engineering life cycle.

The Uncertainty issues during the performance test strategy are virtually unavoidable. The Performance Testing activities can fail in many ways; however, you can prevent most problems by modeling the right workload with appropriate transaction mix & identifying the underlying statistical distribution model of the application under test. Conducting performance tests based on the inputs from the workload model would be the right way of certifying the system performance.

The Objective of our paper is to put forth the workload modeling methodology we adopted in various web applications & to share with the case study, how to characterize the workload for the typical web application in order to simulate the real time usage pattern. The paper also deals with the web log analysis & the metrics to be predicted based on the arrival pattern of the requests in order to perform realistic performance tests to certify the application. The various ways of how the assumptions might fail when it comes to the real time workload are discussed in our paper along with checklist to guide the performance engineer to arrive at the right set of assumptions & techniques.